Signs announcing opposition to a third span over the Chesapeake Bay are already popping up in Queen Anne and Kent County.   Governor Hogan initiated a study of the need, possible location, and potential cost of a new span in 2016, and consultations with potentially affected communities began earlier this year.

There is no question that the Bay Bridge is subject to time-wasting congestion during our evening commutes back to the Eastern Shore and on summer weekends when vacationers head to the beaches.  The number of vehicles trying to cross the bridge is projected to increase by over 30% by 2040, ultimately turning congestion into gridlock.

Relieving that current and projected future congestion is the reason given for building an additional span over the Bay.  But more construction may not be necessary if automated vehicles take over the market as other projections suggest.

I have been working with an organization known as SAFE (Securing America’s Future Energy) on a study of the potential costs and benefits of automated vehicles for the past year.  It was released on June 13. These are some highlights.

New cars already incorporate many new technologies that automate driving tasks:  adaptive cruise control, lane keeping assistance, collision avoidance, and self-parking to name a few.  Many experts and auto companies foresee rapid improvement in these fledgling technologies, to the point that vehicles could drive themselves with little or no need for driver intervention.  Google and Uber are already ordering and testing such vehicles (as news of a recent pedestrian fatality caused by faulty programming of hazard detection and response logic made everyone aware).

In our study, SAFE projected that with favorable technology advances and market conditions, over 90% of passenger miles traveled by 2040 could be in fully automated vehicles.  The capabilities of those vehicles would make it possible to move all the projected traffic over the Bay Bridge in 2040 with zero congestion and no new construction.

Some preliminary projections from the Maryland Department of Transportations set the context.  The Bay Bridge now carries more than 70,000 vehicles a day. MDOT expects that to grow to 92,000 by 2040.  The average traffic volume during weekends in the summer is forecasted to grow to 125,900 vehicles per day by 2040, a 31 percent increase from 2013.

A December 2004 Transportation Needs Report from MDOT analyzed hourly congestion levels, which are what really matter. With the level of traffic projected for 2040, about 5,750 vehicles per hour would attempt to cross the bridge during the weekend peak period of 2 – 5 PM.  Peak weekday traffic going eastbound would be over 4000 vehicles per hour between 4 and 5 PM.

MDOT estimates that when fewer than 2000 vehicles per hour are crossing in the three westbound lanes there is no congestion and when 2000 vehicles per hour are crossing eastbound there will be only occasional slowdowns.

When more than 3000 vehicles per hour attempt to cross in either direction, traffic breaks down and stop and go traffic is the rule.  With that constraint, the levels of traffic projected for 2040 would be catastrophic.

An MDOT study guessed that a new span could cost up to $6.85 billion, and would require other road network upgrades.  With the time required for planning, debating and construction, it is unlikely that would do anything for congestion for at least a decade or possibly longer.

Vehicle automation that we are likely to see could make that new span unnecessary by the time the money is spent.

There are three major ways in which vehicle automation can reduce or eliminate congestion.   The technologies that are required include detection of surrounding vehicles and communication of traffic conditions together with automatic control of speed, braking and lane changing.  These automated capabilities would

1. Allow vehicles to travel safely with much smaller distances between vehicles
2. Eliminate the accordion effects created by lane changing and human reaction times for braking and accelerating
3. Prevent accidents that are the major cause of congestion not caused by inadequate capacity.

Just the first of these benefits, shorter headway, would dramatically increase bridge capacity.

With anything over 2000 vehicles per hour now causing some form of congestion, and potential peak traffic of 5750 per hour during weekends in 2040, the capacity of the bridge would have to be nearly tripled to avoid weekday and weekend congestion.   The worst forms of congestion now appear when traffic exceeds 3000 vehicles per hour, and just to avoid those conditions capacity would have to be doubled.

Building an additional bridge with the same one-way capacity as the current bridge would provide just barely enough additional capacity to accommodate weekday rush hour traffic without congestion, and would still put weekend traffic into stop and go conditions much of the time.

In contrast, cutting the distance between vehicles in half at highway speeds would double the number of vehicles that could cross the bridge with no congestion.  An automated vehicle will be much faster than a human being in braking to avoid a rear-end collision, and communication between vehicles will give it advance warning of traffic conditions far beyond line of sight.  

It is straightforward algebra to determine that if the bridge can handle 2000 vehicles per hour with current distances maintained between vehicles, it could handle 4000 vehicles per hour with half that spacing and 6000 vehicles per hour with one-third the spacing.  From my calculations, the current capacity of the 2-lane eastbound bridge translates into about a 3-second distance between vehicles when traffic is flowing smoothly. If the faster reaction times and ability to observe distant changes in traffic flow characteristic of AVs reduced that to a 1-second distance, the needed tripling of bridge capacity would be achieved.

Thus if AV technology and the share of AVs in the total vehicle fleet progresses to the point that headways can be cut to one-third of the current prevailing distance, there would be no need for a new bridge.  All the needed additional capacity would be provided at no extra cost by the automated vehicle fleet.

What does that imply as a prudent course of action now?  

First, the advancement of AV technology and introduction of automated vehicles should be monitored carefully to determine how introduction of AVs is changing capacity requirements.  For example, traffic studies have found that even if as few as 10% of the vehicles crossing the bridge have automated capabilities, they can smooth out traffic flow.

Second, traffic systems on the bridge should be updated to take advantage of automated capabilities as soon as they appear – for example, creating reserved lanes for vehicles with collision avoidance and automatic cruise control systems once there are enough to utilize such lanes fully.  

Third, traffic management measures like congestion-varying tolls could be used to spread out traffic on the existing bridge, until enough AVs are on the road to increase its peak capacity.  These have been proven on Virginia freeways and would work even better if automated vehicles obtained real time information on tolls.

With this combination of incremental improvements in capacity as AVs become more prevalent and the ultimate increase in capacity from a fully automated fleet, the disruption and expense of a third span over the Chesapeake might be avoided completely.

David Montgomery is retired from a career of teaching, government service and consulting, during which he became internationally recognized as an expert on energy, environmental and climate policy. He has a PhD in economics from Harvard University and also studied economics at Cambridge University and theology at the Catholic University of America, David and his wife Esther live in St Michaels, and he now spends his time in front of the computer writing about economic, political and religious topics and the rest of the day outdoors engaged in politically incorrect activities.